Search Results (2,311)

Valve internal leakage in thermal power stations exhibits a strong concealed nature. If it cannot be discovered and predicted of development trend in time, it will affect the safe and economical operation of plant equipment. This paper proposed an intelligent identification method for valve internal leakage that integrated an Improved Kepler Optimization Algorithm (IKOA) with Support Vector Regression (SVR). The Kepler Optimization Algorithm (KOA) was improved using the Sobol sequence and an adaptive Gaussian mutation strategy to achieve self-optimization of the key parameters in the SVR model. A multi-step sliding cross-validation method was employed to train the model, ultimately yielding the IKOA-SVR intelligent identification model for valve internal leakage quantification. Taking the main steam drain pipe valve as an example, a simulation case validation was carried out. The calculation example used Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) and determination coefficient (R²) as performance evaluation metrics, and compared and analyzed the training and testing dataset using IKOA-SVR, KOA-SVR, Particle Swarm Optimization (PSO)-SVR, Random Search (RS)-SVR, Grid Search (GS)-SVR, and Bayesian Optimization (BO)-SVR methods, respectively. For the testing dataset, the MSE of IKOA-SVR is 0.65, RMSE is 0.81, MAE is 0.49, and MAPE is 0.0043, with the smallest values among the six methods. The R² of IKOA-SVR is 0.9998, with the largest value among the six methods. It indicated that IKOA-SVR can effectively solve problems such as getting stuck in local optima and overfitting during the optimization process. An Out-Of-Distribution (OOD) test was conducted for two scenarios: noise injection and Region-Holdout. The identification performance of all six methods decreased, with IKOA-SVR showing the smallest performance decline. The results show that IKOA-SVR has the strongest generalization ability and robustness, the best effect in improving fitting ability, the smallest identification error, the highest identification accuracy, and results closer to the actual value. The method presented in this paper provides an effective approach to solve the problem of intelligent identification of valve internal leakage in thermal power station. Full article

Modern healthcare and engineering both rely on robust reliability models, where handling censored data effectively translates into longer-lasting devices, improved therapies, and safer environments for society. To address this, we develop a novel inferential framework for the ZLindley (ZL) distribution under the improved adaptive progressive Type-II censoring strategy. The proposed approach unifies the flexibility of the ZL model—capable of representing monotonically increasing hazards—with the efficiency of an adaptive censoring strategy that guarantees experiment termination within pre-specified limits. Both classical and Bayesian methodologies are investigated: Maximum likelihood and log-transformed likelihood estimators are derived alongside their asymptotic confidence intervals, while Bayesian estimation is conducted via gamma priors and Markov chain Monte Carlo methods, yielding Bayes point estimates, credible intervals, and highest posterior density regions. Extensive Monte Carlo simulations are employed to evaluate estimator performance in terms of bias, efficiency, coverage probability, and interval length across diverse censoring designs. Results demonstrate the superiority of Bayesian inference, particularly under informative priors, and highlight the robustness of HPD intervals over traditional asymptotic approaches. To emphasize practical utility, the methodology is applied to real-world reliability datasets from clinical trials on leukemia patients and hydrological measurements from River Styx floods, demonstrating the model’s ability to capture heterogeneity, over-dispersion, and increasing risk profiles. The empirical investigations reveal that the ZLindley distribution consistently provides a better fit than well-known competitors—including Lindley, Weibull, and Gamma models—when applied to real-world case studies from clinical leukemia trials and hydrological systems, highlighting its unmatched flexibility, robustness, and predictive utility for practical reliability modeling. Full article

The aim of this study was to develop a dynamic factorial model for predicting amino acid requirements in Hy-Line Gray laying hens during critical early growth stages (0–84 days), addressing the need for precision feeding in modern poultry production systems. Methods: Four sequential trials were conducted. In Trial 1, growth curves and protein deposition equations were developed based on fortnightly body composition analyses, with parameters evaluated using the Akaike and Bayesian information criteria (AIC and BIC). In Trial 2, the carcass and feather amino acid profiles were characterized via HPLC. And established the amino acid composition patterns of chicken feather protein and carcass protein (AAF and AAC). In Trial 3, maintenance requirements were quantified through nitrogen balance studies, and in Trial 4, amino acid patterns of feather protein (APD) and apparent protein digestibility (ADD) were established using an endogenous indicator method. These datasets were integrated through factorial modeling to predict age-specific nutrient demands. Results: The developed model revealed the following quantitative requirements (g/day) for 18 amino acids across developmental stages: aspartic acid (0.1–0.863), glutamic acid (0.170–1.503), serine (0.143–0.806), arginine (0.165–0.891), glycine (0.258–1.279), threonine (0.095–0.507), proline (0.253–1.207), alanine (0.131–0.718), valine (0.144–0.737), methionine (0.023–0.124), cysteine (0.102–0.682), isoleucine (0.086–0.458), leucine (0.209–1.067), phenylalanine (0.086–0.464), histidine (0.024–0.133), lysine (0.080–0.462), tyrosine (0.050–0.283), and tryptophan (0.011–0.060). The model demonstrated strong predictive validity throughout the 12-week growth period. Conclusion: This integrative approach yielded the first dynamic requirement model for Hy-Line Gray layers during early development. The factorial framework enables precise adjustment of amino acid provisions to match changing physiological needs and has high potential value in optimizing feed efficiency and supporting sustainable layer production practices. Full article

Unconventional reservoirs are critical for future energy supply, but present major challenges for predictions of production due to their ultra-low permeability, strong pressure sensitivity, and non-Darcy flow. Mechanistically grounded physics-based models depend on uncertain parameters derived from laboratory tests or empirical correlations, limiting their field reliability. A data-enhanced variable start-up pressure gradient framework is developed herein, integrating flow physics with physics-informed neural networks (PINNs), surrogate models, and Bayesian optimization. The framework adaptively refines key parameters to represent spatial and temporal variability in reservoir behavior. Validation with field production data shows significantly improved accuracy and robustness compared to baseline physics-based and purely data-driven approaches. Sensitivity and uncertainty analyses confirm the physical consistency of the corrected parameters and the model’s stable predictive performance under perturbations. Comparative results demonstrate that the data-enhanced model outperforms conventional models in accuracy, generalization, and interpretability. This study provides a unified and scalable approach that bridges physics and data, offering a reliable tool for prediction, real-time adaptation, and decision support in unconventional reservoir development. Full article

As global climate change intensifies, understanding drought mechanisms is crucial for managing water resources and agriculture. This study employs the Standardized Precipitation–Actual Evapotranspiration Index (SPAEI), Standardized Runoff Index (SRI), and Standardized Soil Moisture Index (SSMI) to analyze meteorological, hydrological, and agricultural droughts in the lower Songhua River basin. The PLUS model was used to predict future land types, with model accuracy validated using four evaluation metrics. The projected land cover was integrated with CMIP6 data into the SWAT model to simulate future runoff, which was used to calculate future SRI. Drought events were extracted using run theory, while drought occurrence probability and return period were calculated via a Copula-based joint distribution model. Bayesian conditional probability was employed to explore propagation mechanisms. The results indicate a significant increase in multidimensional drought risk, particularly when the cumulative frequency of univariate droughts reaches 25%, 50%, or 75%. Although increased duration and intensity enhance the likelihood of combined droughts, extremely high values cause a decline in joint probability under “OR” and “AND” conditions. Under different climate scenarios, the recurrence intervals of meteorological, hydrological, and agricultural droughts in the lower reaches of the Songhua River exhibit increased sensitivity with severity, demonstrating consistent propagation patterns across the meteorological–hydrological–agricultural system. Meteorological drought was found to propagate to hydrological and agricultural drought within ~6.00 months and ~3.67 months, respectively, with severity amplifying this effect. Propagation thresholds between drought types decreased with increasing intensity. This study combined SWAT and CMIP6 models with PLUS-based land-use scenarios, highlighting that land-use changes significantly influence spatiotemporal drought patterns. Model validation (Kappa = 0.83, OA = 0.92) confirmed robust predictive accuracy. Overall, this study proposes a multidimensional drought risk model integrating Copula and Bayesian networks, offering valuable insights for drought management under climate change. Full article

Crude oil price forecasting is an important task in energy management and storage. In this regard, deep learning has been applied in the literature to generate accurate forecasts. The main purpose of this study is to design an ensemble prediction system based on various deep learning systems. Specifically, in the first stage of our proposed ensemble system, convolutional neural networks (CNNs), long short-term memory networks (LSTMs), bidirectional LSTM (BiLSTM), gated recurrent units (GRUs), bidirectional GRU (BiGRU), and deep feedforward neural networks (DFFNNs) are used as individual predictive systems to predict crude oil prices. Their respective parameters are fine-tuned by Bayesian optimization (BO). In the second stage, forecasts from the previous stage are all weighted by using the sequential least squares programming (SLSQP) algorithm. The standard tree-based ensemble models, namely, extreme gradient boosting (XGBoost) and random forest (RT), are implemented as baseline models. The main findings can be summarized as follows. First, the proposed ensemble system outperforms the individual CNN, LSTM, BiLSTM, GRU, BiGRU, and DFFNN. Second, it outperforms the standard XGBoost and RT models. Governments and policymakers can use these models to design more effective energy policies and better manage supply in fluctuating markets. For investors, improved predictions of price trends present opportunities for strategic investments, reducing risk while maximizing returns in the energy market. Full article

In Nigeria’s northwestern states of Kano, Katsina, and Kaduna, mechanized rice production is an important contributor to household income and rural economic activity, especially amid a rapidly growing population projected to exceed 400 million by 2050. This study investigates the socio-economic insights of mechanized rice farmers and assesses the impact of mechanization on income, seasonal production, government support, and rural poverty alleviation. Data were collected from 125 respondents across 14 local government areas by using structured questionnaires and analyzed through descriptive statistics and hybrid machine learning models. The findings show that revenue generation significantly influences the adoption of mechanized rice farming, while government involvement is limited and largely ineffective. Advanced predictive modeling revealed that hybrid approaches, particularly those combining regression and Artificial Neural Networks with Bayesian Optimization, outperformed traditional models in forecasting rice yield. Key challenges identified include the high cost of equipment and restricted access to subsidized inputs. This study concludes that income from rice sales drives mechanization and that targeted policy interventions are necessary to overcome socio-economic barriers and improve productivity. These findings highlight the dual importance of economic empowerment and technological innovation in advancing sustainable rice production and improving livelihoods in Nigeria’s rice-growing regions. Full article

Background: Liver disease is a growing global health burden. While individual environmental exposures like heavy metals (lead, cadmium, mercury) and behavioral factors such as smoking and alcohol use are known risk factors, their combined impact and the underlying physiological pathways are poorly understood. Allostatic load (AL), a measure of cumulative physiological stress, is a potential mediator or modifier in the relationship between these chronic exposures and liver disease. This study aimed to investigate the joint effects of heavy metals and behavioral exposures on liver health and to examine the mediating role of AL. Methods: This cross-sectional study utilized data from the National Health and Nutrition Examination Survey (NHANES) 2017–2018 cycle. We assessed blood concentrations of the environmental and lifestyle variables in relation to liver biomarkers and the Fatty Liver Index (FLI). Descriptive statistics were used to summarize participant characteristics. Multivariable linear regression and Bayesian Kernel Machine Regression–Causal Mediation Analysis (BKMR-CMA) were used to model combined, nonlinear effects of the exposure–outcome mixture and to evaluate the mediating role of AL. Results: Lead exposure was positively associated with higher AST (β = 0.65, p = 0.04) and GGT (β = 1.99, p = 0.05), while smoking increased GGT (β = 0.79, p = 0.03) and ALP (β = 0.78, p < 0.01). AL independently predicted higher FLI (β = 3.66, p < 0.001). Conclusions: This study highlights that liver health is influenced by the combined effects of environmental pollutants, behaviors, and cumulative biological stress. While lead exposure and smoking were independently linked to liver enzyme elevations, and AL to FLI, mediation by AL was limited, though trends suggest AL may still amplify chronic metabolic pathways leading to liver disease. Full article

Digital twins (DTs), virtual replicas that integrate mechanistic modeling with real-time clinical data, are emerging as powerful tools in healthcare with particular promise in pediatrics, where age-dependent physiology and ethical considerations complicate infectious disease management. This narrative review examines current and potential applications of DTs across antimicrobial stewardship (AMS), diagnostics, vaccine personalization, respiratory support, and system-level preparedness. Evidence indicates that DTs can optimize antimicrobial therapy by simulating pharmacokinetics and pharmacodynamics to support individualized dosing, enable Bayesian therapeutic drug monitoring, and facilitate timely de-escalation. They also help guide intravenous-to-oral switches and treatment durations by integrating host-response markers and microbiological data, reducing unnecessary antibiotic exposure. Diagnostic applications include simulating host–pathogen interactions to improve accuracy, forecasting clinical deterioration to aid in early sepsis recognition, and differentiating between viral and bacterial illness. Immune DTs hold potential for tailoring vaccination schedules and prophylaxis to a child’s unique immune profile, while hospital- and system-level DTs can simulate outbreaks, optimize patient flow, and strengthen surge preparedness. Despite these advances, implementation in routine pediatric care remains limited by challenges such as scarce pediatric datasets, fragmented data infrastructures, complex developmental physiology, ethical concerns, and uncertain regulatory frameworks. Addressing these barriers will require prospective validation, interoperable data systems, and equitable design to ensure fairness and inclusivity. If developed responsibly, DTs could redefine pediatric infectious disease management by shifting practice from reactive and population-based toward proactive, predictive, and personalized care, ultimately improving outcomes while supporting AMS and health system resilience. Full article

Background/Objectives: Intraoperative neurophysiological monitoring (IONM) is used to detect and prevent neurological injury during extramedullary spinal cord tumor (EMSCT) resection, but its diagnostic accuracy lacks systematic validation with recent evidence. This meta-analysis evaluates the performance of somatosensory evoked potentials (SSEPs), transcranial motor evoked potentials (TcMEPs), and multimodal (SSEP + TcMEP) IONM in predicting deficits during EMSCT resections. Methods: Following PRISMA-DTA guidelines, we searched MEDLINE, PubMed, and Ovid (inception to April 2025) for studies on IONM in EMSCT surgeries (PROSPERO: CRD420251047345). Pooled sensitivity, specificity, and reversibility metrics were calculated using bivariate models, with quality assessed via QUADAS-2. Z-test and Bayesian meta-analysis were used for comparisons. Results: Across 20 studies (2672 patients), multimodal IONM showed a log DOR of 4.310 (95% CI: 3.581–5.039) and an AUC of 94.2%, TcMEP monitoring showed a log DOR of 4.367 (95% CI: 3.765–5.127) and an AUC of 92%, while SSEP monitoring showed a log DOR of 3.463 (95% CI: 2.702–4.224) and an AUC of 82%. All modalities demonstrated high specificity (>95%), indicating low false-positive rates. Bayesian analysis revealed >90% probability that TcMEP-based approaches were superior to SSEPs. Reversible TcMEP changes were associated with an 11% (95% CI: 4–24%) postoperative deficit rate, compared to 35% (95% CI: 12–67%) for SSEPs. Conclusions: These findings caution against relying solely on SSEPs and support the use of multimodal IONM strategies, which enhance early detection of impending neurological injury, enable timely surgical interventions, and help prevent permanent neurological damage in EMSCT resections. Although TcMEP and multimodal monitoring showed similar diagnostic accuracy, we continue to recommend multimodal approaches as the current standard of care, pending prospective studies to determine if TcMEP alone can reliably replace multimodal monitoring. Full article

Groundwater level predictive monitoring is necessary to address accelerated aquifer depletion and ensure sustainable management under increasing climatic and anthropogenic pressures. This study uses machine learning approaches to model groundwater level (GWL) dynamics in six observation wells in the Ruataniwha Basin, New Zealand. These models are enhanced with seasonal decomposition techniques. This study uses both static properties and dynamic variables to capture hydrogeological heterogeneity. Random Forest (RF) and Support Vector Machine (SVM), with seasonal decomposition preprocessing, were developed for GWL modelling. The models were trained on 80% of the dataset and tested using the remaining 20% of the data. Model accuracy was assessed using five statistical metrics: mean absolute error (MAE), root mean square error (RMSE), the coefficient of determination (R²), mean absolute percent error (MAPE), and percent bias (PBIAS). Model uncertainty was analyzed using Bayesian Model Averaging combined with the p-factor and d-factor at the 95% confidence level. The results demonstrate that both models delivered strong predictive performance across training, testing, and full period evaluations. However, the RF model demonstrated a marginally superior predictive accuracy by achieving lower errors (MAE: 0.013–0.174; RMSE: 0.04–0.283), better bias control (PBIAS ≈ 0%), and slightly tighter error bounds in most wells. Uncertainty quantification revealed that models provided a minimum p-factor of 0.878, capturing more than 87% of the observed GWL data within the uncertainty bounds. Comparing the results of both models, the RF model has higher p-factor values ranging from 0.878 to 0.976 with precise interval widths (d-factor: 0.436–0.769), indicating its reliability for adaptive groundwater management. Full article

In risk assessment, deriving dermal absorption values is essential for evaluating plant protection products. Applicants submit study data, which authorities assess during the authorisation process. If no data are provided, default values from the European Food Safety Authority 2017 Guidance on dermal absorption (EFSA GD2017) apply. The German Federal Institute for Risk Assessment compiled an updated dermal absorption dataset of 356 more recent human in vitro studies evaluated under to the newest guidance. We applied the same empirical and modelling approaches used to derive default values for concentrates (commercially available product concentrations) and dilutions in different formulation type categories in EFSA GD2017 to the new dataset and compared the resulting values. We also assessed the impact of applying the alternative definition of ‘concentrate’ (>50 g/L) according to SCoPAFF. Default values obtained by analysing the new dataset were considerably lower than current default values, particularly for solids applied in dilutions. The alternative definition of ‘concentrate’ did not have a large impact on default values. Our results suggest that a revision of the default values based on newer studies evaluated under the most current guidance may be warranted. Full article

Background: Arterial hypertension is highly prevalent among patients with chronic kidney disease (CKD), acting both as a cause and consequence of declining kidney function, and significantly increasing cardiovascular risk. Among modifiable risk factors, diet—particularly excessive sodium intake—plays a central role in the prevention and personalized management of CKD. Methods: This study aimed to develop an innovative, digitally accessible tool to estimate sodium intake in stages 3 to 5 CKD patients, using 24-h urinary sodium excretion as the reference standard. Results: Twenty-five clinical, biological, therapeutic, and dietary variables were collected from 493 patients followed across 6 French centers. A probabilistic Tree-Augmented Naive Bayes model was used to develop the tool based on the 15 most informative variables. The model demonstrated an internal accuracy of 71%, indicating that predicted and observed sodium intake categories matched in 71% of cases. Conclusions: This AI-based prediction model offers a promising clinical tool to estimate daily sodium intake in patients with stages 3 to 5 CKD. However, external validation using independent national and international datasets is essential to establish its robustness and generalizability prior to implementation in routine clinical practice. Full article

Predicting subsoil properties accurately is important for engineering tasks like construction, land development, and environmental management. However, traditional approaches that use borehole data often face challenges because the data is sparse and unevenly spread, which can cause uncertainty in understanding the subsurface. This study introduces a novel geostatistical framework employing Empirical Bayesian Kriging 3D (EBK3D) within a Geographic Information System (GIS), which was developed to construct three-dimensional lithological models. The framework was applied to 265 boreholes from the Queen Mary Reservoir in London. ArcGIS Pro was used to interpolate lithology layers using EBK3D, resulting in voxel-based models that represent both horizontal and vertical lithological variations. Model validation was performed with an independent dataset comprising 30% of the boreholes. The results demonstrated high predictive accuracy for layer elevations (Pearson’s r = 0.99, MAE = 0.31 m). The model achieved 100% accuracy in predicting borehole stratigraphy in homogenous zones and correctly identified 77% of lithological layers in heterogeneous zones. In complex regions, the model accurately predicted the whole borehole in 49% of cases. This framework provides a reliable, repeatable, and cost-effective method for three-dimensional subsurface characterization, enhancing traditional approaches by automating uncertainty quantification and capturing both vertical and horizontal variability. Full article

Digital soil mapping has become increasingly important for large-scale soil organic carbon (SOC) assessments, yet the choice of accuracy assessment method significantly influences model performance interpretation. This study investigates the impact of cross-validation fold numbers on accuracy metrics and compares cross-validation with split-sample validation approaches in national-scale SOC mapping. Five machine learning algorithms (Random Forest, Cubist, Support Vector Regression, Bayesian Regularized Neural Networks, and ensemble modeling) were evaluated to predict SOC content across France (539,661 km²) and Czechia (78,873 km²) using 2731 and 445 soil samples, respectively. Environmental covariates included satellite imagery (Sentinel-1, Sentinel-2, and MODIS), climate data (CHELSA), and topographic variables. Four cross-validation approaches (k = 2, 4, 5, 10) were utilized with 100 repetitions each and the results were compared with the existing literature using both cross-validation and split-sample methods. Ensemble models consistently produced the highest prediction accuracy and lowest variance per fold across all validation approaches. Higher fold numbers (k = 10) also produced higher accuracy estimates compared to lower folds (k = 2) and had the greatest value ranges of accuracy assessment metrics. This confirmed the observations from previous studies, in which split-sample validation reported higher R² values (0.10–0.90) compared to cross-validation studies (0.03–0.68), suggesting a strong effect of randomness in training and test data split in the split-sample approach. This suggests that k-fold cross-validation should preferably be used in reporting prediction accuracy in similar studies, with the split-sample approach being strongly affected by value properties from training and test data from particular splits used for validation. Full article